Draft SOO ACADM Rev 1.pdf
Posted: Aug. 26, 2026
• Type: .pdf
• Size: 0.32MB
Opportunity Assistant
Loading
Overview
Related Opportunity
Related Agency
Summary
This Draft Statement of Objectives, issued by the Defense Health Agency (DHA), defines a prototype requirement for AI/ML-enabled autonomous clinical encounter coding and automated appeals and denials management integrated with MHS GENESIS.
DHA seeks capabilities that can process approximately 143,000 encounters per day—or 4.35 million per month—with a maximum 24-hour turnaround, assign complete billing codes and modifiers without human intervention, maintain source-documentation audit trails and explainability, identify and prioritize denied claims, generate appeal artifacts, assess denial root causes and payer performance, and prevent coding-related denials before claim submission.
The effort is structured around two capabilities and four gated milestones.
For Autonomous Coding, the vendor must:
- Complete CAC credentialing and IATT cybersecurity requirements.
- Achieve at least 60% autonomous coding coverage of approved pre-production test volume with at least 95% code-level accuracy.
- Obtain an ATO-C for an IL4+-authorized hosting environment.
- In production, autonomously code at least 80% of total in-scope volume with at least 95% accuracy.
- Maintain average processing under five minutes and ensure no processing occurs later than 24 hours.
- Provide a 100%-reliable human-routing fallback for at least 90 days.
Testing is prioritized across outpatient and clinic specialties, behavioral health, laboratory, emergency medicine, outpatient radiology, immunizations, primary care, specialty care, and day surgery.
For Appeals and Denials Management, the solution must:
- Ingest 100% of active, open, and aged accounts-receivable records.
- Achieve at least a 90% correlation match against a human-audited sample of 500 claims.
- Map all identified denials to standardized root causes.
- Reduce coding-related denials by at least 20% while maintaining or increasing net third-party collections per encounter.
- Generate appeal artifacts for at least 50% of prioritized recoverable claims, with credit based on claims subsequently collected or overturned.
Collaborative vendors, if separate providers are selected, must deliver unified dashboards, coordinated machine-learning insights, secure HIPAA- and DoD-compliant data exchange, and eventual integration with approved MHS GENESIS interfaces without modifying core functionality.
Proprietary model weights, source code, training algorithms, and similar technical data need not be shared and must be protected through an ACA or reciprocal NDA.
Financial dispositions such as write-offs, adjustments, settlements, or claim-value determinations require human approval. All automated actions must have FFMIA-aligned audit trails.
Deliverables include:
- Monthly progress reports due by the 10th.
- Quarterly in-process reviews.
- Model validation reports after each test.
- Monthly invoice support based on successfully encoded encounters and complexity tiers.
- A final project report within 90 days of prototype completion.
The schedule is driven by authorization and audit requirements rather than an accelerated fixed date. Milestone 1 is targeted within 90 days of the anticipated August 2026 award, pre-production testing is targeted within 90 days after IATT, production entry is contingent upon ATO-C, and the estimated period of performance is 13–16 months.
Both capabilities are targeted for production integration and validation by June 30, 2027.
DHA seeks capabilities that can process approximately 143,000 encounters per day—or 4.35 million per month—with a maximum 24-hour turnaround, assign complete billing codes and modifiers without human intervention, maintain source-documentation audit trails and explainability, identify and prioritize denied claims, generate appeal artifacts, assess denial root causes and payer performance, and prevent coding-related denials before claim submission.
The effort is structured around two capabilities and four gated milestones.
For Autonomous Coding, the vendor must:
- Complete CAC credentialing and IATT cybersecurity requirements.
- Achieve at least 60% autonomous coding coverage of approved pre-production test volume with at least 95% code-level accuracy.
- Obtain an ATO-C for an IL4+-authorized hosting environment.
- In production, autonomously code at least 80% of total in-scope volume with at least 95% accuracy.
- Maintain average processing under five minutes and ensure no processing occurs later than 24 hours.
- Provide a 100%-reliable human-routing fallback for at least 90 days.
Testing is prioritized across outpatient and clinic specialties, behavioral health, laboratory, emergency medicine, outpatient radiology, immunizations, primary care, specialty care, and day surgery.
For Appeals and Denials Management, the solution must:
- Ingest 100% of active, open, and aged accounts-receivable records.
- Achieve at least a 90% correlation match against a human-audited sample of 500 claims.
- Map all identified denials to standardized root causes.
- Reduce coding-related denials by at least 20% while maintaining or increasing net third-party collections per encounter.
- Generate appeal artifacts for at least 50% of prioritized recoverable claims, with credit based on claims subsequently collected or overturned.
Collaborative vendors, if separate providers are selected, must deliver unified dashboards, coordinated machine-learning insights, secure HIPAA- and DoD-compliant data exchange, and eventual integration with approved MHS GENESIS interfaces without modifying core functionality.
Proprietary model weights, source code, training algorithms, and similar technical data need not be shared and must be protected through an ACA or reciprocal NDA.
Financial dispositions such as write-offs, adjustments, settlements, or claim-value determinations require human approval. All automated actions must have FFMIA-aligned audit trails.
Deliverables include:
- Monthly progress reports due by the 10th.
- Quarterly in-process reviews.
- Model validation reports after each test.
- Monthly invoice support based on successfully encoded encounters and complexity tiers.
- A final project report within 90 days of prototype completion.
The schedule is driven by authorization and audit requirements rather than an accelerated fixed date. Milestone 1 is targeted within 90 days of the anticipated August 2026 award, pre-production testing is targeted within 90 days after IATT, production entry is contingent upon ATO-C, and the estimated period of performance is 13–16 months.
Both capabilities are targeted for production integration and validation by June 30, 2027.
Show All